AI Engineer - LLMs, RAG and agents

Modders Consulting; Singapore; Hybrid; SGD 87,000 - 131,000; Full time

Job summary

AI platforms and assistants built on language models are developed for production, focusing on RAG pipelines, agent orchestration and business APIs. The work covers concrete topics: document ingestion, agents, and APIs with production deployment in mind.

The role emphasizes building end-to-end pipelines, integration with existing systems, and optimizing answer quality, data structuring and performance, with a strong focus on ML ops practices and vector databases.

Qualifications

• Experience in AI engineering with language models and retrieval augmented generation.
• Strong Python and orchestration framework skills (e.g., LangChain).
• Ability to build end-to-end pipelines and integrate into existing systems.
• Understanding of answer quality, data structuring and performance concerns.
• MLOps sensibility and vector databases preferred.

Responsibilities

• Designing and building specialised AI assistants, with multi-step workflow and agent orchestration.
• Designing retrieval augmented generation pipelines: document ingestion, chunking, embeddings, ranking and retrieval.
• Continuously improving answer quality and running tool/e model evaluation.
• Building APIs that expose assistants to business applications.
• Monitoring usage and keeping performance and inference costs under control.

Skills

AI engineering

Python

LangChain

LLM/RAG concepts

Pipeline design

Performance optimization

Tools

LangGraph

RAG pipelines

FastAPI

PostgreSQL/pgvector

Docker

Kubernetes

Azure AI

CI/CD

Job description

AI platforms and assistants built on language models and designed for production: RAG pipelines, agent orchestration, business APIs. No demoware.

• Paris / Île-de-France
• Permanent contract
• 1+ year of experience
• AI and data platforms

AI systems that are actually used

We work on AI platform and assistant projects built on language models, with industrialisation and production deployment in mind. The topics are concrete: document ingestion, agents, business-facing APIs.

The quality of an AI system comes from the data, the pipeline and the architecture, not from the prompt alone. That conviction shapes how we approach these projects.

Building pipelines and agents that hold up in production

• Designing and building specialised AI assistants, with multi-step workflow and agent orchestration.
• Designing retrieval augmented generation pipelines: document ingestion, chunking, embeddings, ranking and retrieval.
• Continuously improving answer quality and running tooled model evaluation.
• Building APIs that expose assistants to business applications.
• Monitoring usage and keeping performance and inference costs under control.

Technical environment

• Python
• LangChain
• LangGraph
• RAG
• FastAPI
• PostgreSQL / pgvector
• Docker
• Kubernetes
• Azure AI
• CI/CD

Who we are looking for

• Experience in AI engineering, on language model and retrieval augmented generation work.
• Good command of Python and of an orchestration framework such as LangChain.
• Ability to build a complete pipeline and integrate it cleanly into an existing system.
• Understanding of answer quality, data structuring and performance concerns.
• An MLOps or LLMOps sensibility and experience with vector databases are appreciated.

This role is not a fit for a purely prompt-oriented or demo-oriented profile.

Why this role is worth your time

• The projects are industrialised and meant to be used, not presented.
• The technical content is substantial: architecture, pipeline, integration.
• The domain is still being built: your decisions shape the platform.
• You get the full technical context before any interview with the client.